J. Phys. Chem. 1988, 92, 563-564 at 1400-1700 K, causing polycyclic aromatic molecules to adopt pentagonal rings in preference to leaving dangling bonds. Implicit in their assumption is the low concentration of hydrogen at those conditions. However, hydrogen in various chemical forms (H, Ha, H,O, ...) is most abundant in hydrocarbon flames. The issue is not the hydrogen content but the kinetics and thermodynamics of the reaction network discussed above. In terms of the product soot, the spiralling cluster model allows for a low H / C ratio. Can we reconcile the benzenoid model for soot with the atomic H / C ratios of 0.1-0.2 typically found in "old" soots?5 A hexagonal lattice crystallite of edge 2 nm will have a periphery of 8 nm and an area of 3.46 nm2. Assuming the hydrogen edge density of acenes 245 pm/H and the graphitic carbon density 0.0260 nm2/C, we have H / C = 0.245 (32.6 H/133 C). If, however, one benzenoid array (Le., continuous covalent network) were composed of two crystallites, we have H / C = 0.18, and so on. Since microscopy shows that crystallite sizes, as determined by diffraction, are smaller than molecular arrays seen in microscopy, we see that the benzenoid array model can reconcile diffraction, microscopy, and microanalytical findings. Many models have been proposed which link benzenoid arrays to spherical m o r p h ~ l o g y ~and , ~ ,thus ~ ~ there is no compelling need to invoke spherical clusters to account for the structure of soot on a molecular, crystallite, or particle level. Although C6,-type structures proposed by ZOHLCKS may indeed be formed under certain conditions's4 including flames$5 the arguments presented in this paper, both kinetic and structural, indicate that the importance of these structures is very unlikely in soot formation. Acknowledgment. The work at Penn State was supported by the Aerothermochemistry Branch of NASA-Lewis Research Center, Grants NAG 3-477 and NAG 3-668. Registry No. C, 7440-44-0. (43) Donnet, J. B. Carbon 1982, 20, 266-282. (44) Iiiima. S.J. Phvs. Chem. 1987. 91. 3466-3467. (45) Gerhardt, Ph.; Loffler, S.;Homann; K. H.Chem. Phys. Lett. 1987, 137, 306-310.
Fuel Science Program Department of Materials Science and Engineering Pennsylvania State University University Park, Pennsylvania 16802 Exxon Corporate Research Annandale, New Jersey 08801
Michael Frenklacb*
Lawrence B. Ebert
Received: July 16, 1987; In Final Form: September 10, 1987
Ultraviolet Spectrum of Chlorine Perchlorate Sir: Chlorine perchlorate (C10C103) was synthesized in 1970 by Schack and Pilipovich,' who determined its principal physical properties including the infrared spectrum. This substance is also produced with good yield during the chlorine dioxide photolysis.24 Gaseous chlorine perchlorate in Pyrex or quartz containers at room temperature shows some instability and also proves to be very sensitive to traces of humidity. Consequently, its UV (1) Schack, C. J.; Pilipovich, D. Inorg. Chem. 1970, 9, 1387. Christe, K. 0.;Schack, C. J.; Curtis, E. C. Inorg. Chem. 1971, 10, 1589. (2) Jubert, A. H.; Sicre, J. E.;Schumacher, H. J. Presented at the 2nd Congreso Argentino de Fisicoquimica, Carlos Paz, C6rdoba, Argentina, Sept 1-5, 1980; paper 83. (3). Schell-Sorokin, A. J.; Bethune, D. S.; Lanckard, J. R.; Loy,M. M. T.; Sorokin, P. P. J. Phys. Chem. 1982, 86, 4653. (4) Barton, R. A.; Cox, R. A. Wallington, T. J. J. Chem. SOC.,Faraday Trans. 1 1984, 80,2731.
0022-3654/88/2092-0563$01.50/0
563
CI OCIO,
b
Q
I,,,,,\-
,
~
2 50
300
350
WAVELENGTH l n m l
Figure 1. Ultraviolet spectrum of chlorine perchlorate.
spectrum, free from strong UV absorbers like C102 and ClO,, is difficult to obtain. This spectrum was unknown, and its knowledge was necessary for us since large amounts of ClOClO, were produced in our experiments on the Cl02 photoly~is.~After careful handling we were able to obtain this spectrum, and as this information is necessary for a correct interpretation of some chlorine oxide reactions$,6 it seems advisable to publish it. In a quartz cylinder reactor (4 = 5 cm; I = 5 cm) 100 Torr of C102 a t 30 OC was photolyzed (A = 436 nm) until it was practically consumed. (The C10C103 yield is approximately 25%.) The low vapor pressure products, essentially C103and the dimer C1206,were eliminated by fractional condensation at -35 OC in a U-form low-volume trap connected between the reactor and the stopcock. The volatile chlorinated compounds Cl,, C102, and C10C103 were condensed in a Pyrex trap at liquid air temperature. In this trap the products of six other batches were put together. Purification was made by eliminating firstly C1, at -100 OC and secondly the small amount of C102 at -60 OC. This last part of the distillation was carefully carried out and controlled by UV spectrophotometry until practically no C102bands were detected. Large amounts of CIOCIO, were lost (of the order of 50%) because of the low vapor pressure ratio [p,(C102)/pv(C10C10,)], oc = 8.7/2.8. Pure C10C103, as a pale greenish liquid, remains at the bottom of the trap. The UV spectrum (Figure 1) was obtained at 30 "C with a Cary 14 spectrophotometer (10-cm quartz cell length). The line represents the average of four runs obtained with different samples. The ClOClO, gas pressures were 1.9, 4.3, 5.1, and 7.7 Torr. At = 234 nm, the values for In (Io/I)are 0.624, the wavelength ,A, 1.184, 1.444, and 2.372, respectively, and the calculated (Beer's cm2 law) absorption cross section is 6234nm = (94 f 7) X molecule-'. In the wavelength range 210-280 nm the error is 10% (maximum-minimum values), and it is caused by the pressure measurement uncertainty. The spectra show that CIOClO, is practically free from ClO,, ClO,, and C1207(maximum limits of impurities OS%, 1.5%, and 0.1%, respectively). However, since beyond 280 nm the absorption of the residual C10, has been deducted, the error naturally increased, and at wavelengths larger than 300 nm the reported cross section values only have a qualitative meaning. The purity of the CIOClO, was checked by IR spectrophotometry and vapor pressure measurements.' The results agree with the quoted values.' Because of the C10C103 instability, the spectra were recorded as fast as possible. For instance, with a sample of 7.7 Torr, decomposition products (mainly C10,) are noticeable after a few minutes and in an hour 8% is decomposed. Our results confirm, to a certain degree, the assignment that
( 5 ) Lopez, M. I.; Sicre, J. E.; Schumacher, H. J., to be published. (6) Molina, L. T.; Molina, M. J. J. Phys. Chem. 1987, 91, 433.
0 1988 American Chemical Society
564
The Journal of Physical Chemistry, Vol. 92, No. 2, 1988
Barton et aL4 have done for the absorption of products of C10, photolysis after subtraction of C102and the transient species C10 and CI2O2. They reported a maximum at 250-260 nm. It seems that a much better coincidence with our value could now be achieved by applying for the absorption of CI2O2the data recently 8 X lo-'* obtained by Molina.6 Nevertheless, the value u,,, cm2 molecule-' appears to be rather high compared with our directly obtained value.
Comments Registry No. CIOCIO,, 27218-16-2.
Instituto de Investigaciones Fisicoqujmicas Tedricas y Aplicadas Sucursal4, Casilla de Correo 16, 1900 La Plaza, Argentina
M. I. Mpez J. E. Sicre*
Received: July 7. 1987; In Final Form: September 8, 1987